Departmental Papers (CIS)

Date of this Version

4-2021

Document Type

Conference Paper

Comments

24th International Conference on Artificial Intelligence and Statistics (AISTATS) 2021, April 13-15 2021, Virtual. PMLR: Volume 130

Abstract

Providing reliable model uncertainty estimates is imperative to enabling robust decision making by autonomous agents and humans alike. While recently there have been significant advances in confidence calibration for trained models, examples with poor calibration persist in most calibrated models. Consequently, multiple techniques have been proposed that leverage label-invariant transformations of the input (i.e., an input manifold) to improve worst-case confidence calibration. However, manifold-based confidence calibration techniques generally do not scale and/or require expensive retraining when applied to models with large input spaces (e.g., ImageNet). In this paper, we present the recursive lossy label-invariant calibration (ReCal) technique that leverages label-invariant transformations of the input that induce a loss of discriminatory information to recursively group (and calibrate) inputs – without requiring model retraining. We show that ReCal outperforms other calibration methods on multiple datasets, especially, on large-scale datasets such as ImageNet.

Subject Area

CPS Machine Learning

Publication Source

International Conference on Artificial Intelligence and Statistics (AISTATS 2021)

Volume

130

Share

COinS
 

Date Posted: 20 October 2021

This document has been peer reviewed.